Bull. ent. Res. 67, 185-204 185 Published 1977 ..

Migrations of spp. ( : ) related to movements OP the Inter-Tropical Convergence Zone in West Africa

DOMINIQ~DUVIARD * Laboratoire d'EntotnoZogie Agricole, Centre Orstom d'Adiopodoumé, B.P. V51, A bidjan, Zvory Coast

Abstract The possibilities of migrations in the West African species of Dysdercus are discussed and a hypothesis of long-range migrations asso- ciated \with the Inter-TTopical Convergence Zone and its wind systems is proposed. Catches of adult Dysdercus in four light-traps distributed from south to north of the Ivory Coast showed that the phenology of assumed migratory activity in D. voelkeri Schmidt differs with latitude and may be correlated with particular types of weather; stainer migrations taking place during the warm, wet and sunny part of the year. The whole life cycle of the as well as their flight activity occur under these climatic conditions, prevailing in a belt of 600-900 km width, situated immediately to the south of the Inter-Tropical Front. Colonisation of newly available habitats is thus only possible when climatic factors allow: (i), migratory flight activity and (ii), survival in the colonised area. A close examination of the timing of both migrations in the two main species, D. voelkeri and D. mehoderes Karsch, and of annual movements of the I.T.F. leads to the only logical hypothesis that the transportation of migrating insects is effected by atmospheric convergence, prevailing wind currents and air mass displacements.

Introduction Migration of adults of a new generation from one breeding site to another is an important feature of the biology of many species (Southwood, 1960; Johnson, 1969; Bowden, 1973; Dingle, 1974) and migrants must therefore be considered as active colonisers of every potential habitat, and not just as individuals leaving an unsuitable environment (Dingle, 1972). This recent concept of migratory behaviour modifies drastically the concept of colonisation of new habitats, such as a field, for example, by pests. Since the early work of Ballard 8c Evans (1928), Golding (1928), Bebbington & Allan (1936), and others, reviewed by Pearson (1958), little field work has been done on migrations of cotton stainers (Dysdercus spp.). The accent in studies on migration has been on physiological aspects in the work of Edwards (1969n, 6, 1970), Dingle & Arora (1973) and Dingle (1972, 1974) and field work has been restricted, until recently, to occasional observations by light-trapping (Gibbs & Leston, 1970; Bowden, 1973) or to the classical type of cotton field survey (Pierrard, 1972). Curiously, the idea that stainers * Present address : Laboratoire d'Entomologie Agricole, Services.Scientifiques Centraux de - l'O.R.S.T.O.M., 70-74 Route d'Aulnay, 93140 BONDY, France. ~I (L 2283) A

i- . .L 186 DOMINIQUE DUVIARD

are common migrants is based on old observations which have not attracted much interest from field workers, possibly because these insects can be controlled easily with pesticides. In the laboratory, the life of adult stainers, at least of the females, may be separated into two distinct phases: (i), a migratory period of enhanced locomotory activity by young, non-feeding, sexually immature imagos with well-developed flight muscles; (ii), a subsequent period during which feeding and mating induce flight muscle histolysis in the females and in some males (which are then unable to fly), and also induce the maturation of ovaries in females. These two phases also exist in the field, where differences in behaviour have been observed (Duviard, 1972). During the migratory period, the adult stainers make nocturnal flights, which end with the discovery of the host plant, followed by the reproductive and sedentary period. Such migratory flights are dependant on climatic conditions (Duviard, 1973) and bugs in the migratory state will fly in the correct conditions of sunshine, sir temperature and atmospheric humidity. Annual observed variations in this migratory phenology led to a hypothesis of a link between migrations and movements of the Inter-Tropical Convergence Zone (I.T.C.Z.). Such an hypothesis was fi~st,but padally, proposed by Golding (1928) and Bowden (1973) later developed this idea with the aid of some results from the work reported here. In the present paper too, the argument will be essentially phenological, strengthened by direct evidence of seasonal displacements, distributions and incidence of colour forms.

Review of available evidence The precariousness of habit ats Southwood (1962) discussed the importance of temporary habitats in the evolution of migratory behaviour of insects. Although useful and fitting the case of cotton stainers, this idea is not wholly acceptable in the Tropics (Bowden, 1964). The temporary nature of the habitats of cotton stainers is indicated in the early works on Dysdercus spp., summarised by Pearson (1958) and its importance is assumed in the recent work , of Dingle & Arora (1973) and Dingle (1974), although no field experiments were presented by these authors. -

IA B 'DlSTRlBUTlON OF IN SAVANNA ISTRlBUTlON OF MALVALES IN RAIN FOREST Adansonlo diglloto - Bomhox buonopozense, Ceiba pentandro. Bombax costatum ._. .Slerculio tropacanlho Sterculia setigero _.._ Fig. I.-Distnbution of the main malvaceous host-plants of Dysdercus spp. in the savanna (A) and forest (B) areas. Within limits of species, lightly dotted areas represent presence of plants and heavily-dotted areas show their optimal range. MIGRATIONS OE DYSDERCUS IN WEST AFRICA 187

Two types of factor are responsible for the precariousness of the habitats of Dysdercus spp. : Food factors.-The host plants (either malvaceous hosts or occasional graminaceous hosts in the case of D. voelkeri Schmidt) provide fruits @bolls) and seeds that are necessary for the development .of Dysderciis (Pearson, 1958). These sources of food are available during a part of the year only. Each host species releases its seeds during a shont period of time (a few weeks to a few months) in a given area and the different plant hosts do not present a similar seeding phenology although, for a given host plant, sharp variations may be observed with changes of latitude. Although the seeding periods of the different host plants are not wholly synchronous, it does noi seem that, when added to each other, they would facilitate the survival of the insects throughout the year at one locality (Whitfield, 1933; Duviard, 1977). Host plants are scattered in the whole non-steppe zone of West Africa (for the names of vegetation zones, see Keay, 1959), ranging from dry Sudan savanna to evergreen rain forest (Fig. 1). Climatic factors.-The sensitivity Of Dysdercns spp. to tempenature and humidity affecting the survival of eggs 'and young nymphs (Pearson, 1958) as well as the activity of nymphs and adult stages (Youdeowei, 1967; Duviard, 1973), is well known. We have previously shown that, in the case of D. voelkeri, favourable climatic conditions are generally encountered together with available food (Duviard, 1973), but this is not a perfect synchrony. During the dry season, the southward movement of the I.T.C.Z. may, with the establishment of the harmattan (a dry warm wind from the Sahara), restrict development of cotton stainers for some time in the savanna habitets, although food may (beavailable. Dryness and high temperature associated with this wind first stop the activiiy of adult and immature stages and eventually kill the eggs and nymphs (Duviard, 1973). In the min forest $belt,areas with much available food will {become unsuitable when rain is heavy. High mortality then occurs due to drowning of colonies of nymphs by rainfall (Galichet, 1956) and to a lethal dungal disease associated with high humidity (Duviard, 1977). Thus, the habitats of Dysdercus (i.e., fruiting host plants and the area of ground where seeds are dispersed), either do or do not exist. If they exist, they may be colonised or not by the insects, depending on climatic conditions.

Sudden immigrations Authors working on D. voelkeri (the species referred to '51s D. sziperstitiuszis (F.) until the revision by Pierrard, 1967) in West or Central Africa have observed that, in a given area where the species was completely absent for several months, the appeared suddenly in large numbers. It is significant that the term 'immigration', used by Golding (192S), has since been used often in this context. Examination of data reveals two distinct types of immigration : Immigration occurring with the omet of the dry semon.-This type has been more often described fur it is observed in the cotton-growing area during the period when cotton forms bolls land seeds which may be colonised by stainers (Golding, 1928, at Ibadan, Nigeria; Pierrard, 1972, at Bambari, Central African Republic; Gibbs & Leston, 1970, at Tafo, Ghana; Bowden, 1973, at Kwadaso, Ghana; Duviard, 1973, at Foro-foro, Ivory Coast). This immigration follows 'il total disappearance of the species during ithe preceding rainy season, and is observed only in the southern part of the geographicaI area of D. voelkeri. Immigration occurring with tlze onset of reins.-This has seldom been described, for it happens during the absence of cotton crops from the fields (Whitfield, 1933, in the Sudan; Golding, 1928, at Kano, Nigeria: Duviard, this paper, at Ferkessedougou, Ivory Coast; (Issa Kone, personal communication, at Bamako, Mali). It follows a complete disappearance of rhe species during the preceding dry season and is observed only in the northern part of the geographicalarea of D. voelkeri. 188 L70kllhlQUE DUVIAF.1)

Thus, in the whole area exploited by D. i90elXeri in West Africa, which is limited in the north by the SOO mm isohyet (according to Whitfield, 1933), there appear to be two distinct areas : a southern one, roughly Guinean and sub-Sudanian, where insects immigrate suddenly with the dry season and disappear with the rains and a northem one, roughly Sudanian, where insects immigrate with the rains and disappear during the dry season. However, Leroux (1972) has shown that dry season and wet season climates are not uniform throughout West Africa and this is referred to again below. Colour forms of D. voelkeri DeLattre (1950) writes that several authors have outlined the seasonal variations of colour observed in D. voelkeri (as D. sliperstitiosus) and have spoken of a nzigrating form moving towards cotton fields and a return form, moving towards wild host plants at the end of the cotton-growing season. The presence of pterin pigments in Dysdercus (Berridge, 1965; Halfenberg & Stein, 1971) and the experiments of Pierrard (1972) show that food does not affect colour, and the cyclic character of the succession of colour forms seems to be linked with climatic conditions (Duviard, 1973). The immigration occurring at the start of the dry season is more or less completely performed by ’ yellow ’ insects (Golding, 1928; Pierrard, 1972; Duviard, 1973). The one to three following generations are all of ‘ red ’ insects and the ’ orange ’ form appears with the return of the rains (March in Foro-foro; Duviard, 1973) but popula- tions of this form quickly disappear from southern and central Ivory Coast. The immigration occurring in Ferkessedougou and in the Bamako area with the onset of rains in March-April, is by ‘ orange ’ insects. The population breeding in the Sudanian area during the rains, as well as the very rare individuals caught at that period of the year in the Guinean belt, are always ‘ yellow ’. Thus it seems that. the yellow form, which appears in the Sudanian belt during the rains, is the immigrant population invading the Guinean belt as the dry season starts, giving rise to a local red form which breeds there until the return of the rains, when it is replaced by an orange form which immigrates into the Sudanian area. It is sometimes difficult to distinguish red and orange specimens, but the yellow form is perfectly distinct. The mixture, in different proportions, of the pterins known from these insects (Berridge, 1965; Halfenbeg & Stein, 1971) is probably responsible for these different patterns. The relative importance of the three colour forms, both in time and space, is shown in Table I. The Ted and orange forms are present every- where, but $he yellow one has not been observed at Lamto or Adiopodoume (see below for description of the experiment). Where it does occur, the yellow form is dominant during the rainy period, while the red form is typical of the dry season, being partly replaced by the orange phase with the beginning of the rains.

TABLEI. Percentage of the three diflerent colour forins of D. voelkeri caught in light-traps at four locafions in the Ivory Coast during 1973 JFMAMJJASOND 100 100 99.2 93.2 43 50 15 7.5 18 100 - 0.8 6.8 21 - - 14.5 35 - - 36 50 85 I8 41 - - - 100 99.8 99.8 100 - - - - 6-- Foro-foro {%&!e - - 0.2 0.2 - - - 50 - 4-- lYellow ------100 50 - 90 - - Red 100 89 100 100 100 ------Lamto Orange - I]----- 100. - - - - Yellow------Red 100 100 100 78.5 100 - - - 100 - 100 92 Orange - - - 21.5------8 Yellow ------_____

I ì MIGRATIONS OF DYSDERCUS IN WEST AFRICA 189

The phenology of fright activity of D. voelkeri at different latitzrdes The few published observations on ,the periodicity of flight of Dysderciis in West Africa, all obtained by light-trapping, are: Tafo, Ghana, 6"13'N, O"22W (Gibbs & Leston, 1970). Kwadaso, Ghana, 6"42'N, O"38W (Bowden, 1973). Foro-foro, Ivory Coast, 7"58'N, 5"OlW (Duviard, 1973). Some data have also been obtained by counting or by other observations at the following localities : Ibadan, Nigeria, 7"25'N, 3 "55'E } (Golding, 1928). Kano.-- Nigeria. 12"N. 8"35'E Adiopodoumé; Ivory'Coast, 5" lYN, 4" 08'W (Galichet, 1956). Accra, Ghana, 4" 15", Oo15W.(Fuseini, 1972). These observations suggest that: (i) The immigration linked with the onset of the dry season occurs earlier at higher latitudes : mid-October at Foro-foro; beginning of November at Kwadaso; mid- December at Tafo. (ii} The lower the latitude, the earlier the disappearance of the insects: April at Adiopodoumé; May-June at Tafo and Kwadaso; June at Foro-foro. The disappearance of D. voelkeri occurs later (June-July) in the Accra area and seems to be correlated with the drier climate of this part of ,the Guinean belt. However, these comparisons come from observations in different years. It seems, as Golding (1928) and Bowden (personal communication) have suggested, and as we have shown in the particular case of Foro-foro, that flight is linked with ,well- defined types of weather, related to movements of the I.T.C.Z. At all localities cited here, which are, except for Kano, in the Sudanian area where population outbreaks occur during the rainy season, migratory flight seems to occur during the least rainy part of the year (interseason, long dry season-see below). The migratory flights are dependent on climatic conditions, as shown at Foro-foro (Duviard, 1973). The interruption of activity, correlated with the harmattan, becomes less as the importance of this seasonal wind decreases with lower latitudes. It is therefore suggested that D. voelkeri is nomadic arid that $he habitats it exploits are always temporary ones. Migratory activity, controlled by both internal factors (Duviard, 1972; Dingle & Arora, 1973; Dingle, 1974) and external factors (Duviard, 1973) allows the regular colonisation of these seasonal habitats. But several facts raise the question of the geographical importance and extent of migration. Are stainers restricted to cyclic run wihin limited areas, as suggested by ear!y authors (see Pearson, 1958)? In a given area, the total seasonal importance of the insects does not fit such an explanation. The various ecological zones colonised by the species. with their considerable climatic differences, cannot be exploited simultaneously by the bugs, if one considers their temperature and humidity require- ments. The alternation, in space and time, of sudden immigrations and departures; the cyclic turnover of different colour forms; the marked phenological differences between localities, are all presumptions of long-range migrations. It was thus necessary to add to the disparate data in the literature to test this hypothesis; to define the conditions for migrations, and to understand their importance in the general context of West Africa.

Experimenfs in the Ivory Coast Light-traps Southwood (1961), Gibbs & Leston (1970), Duviard (1972, 1973) and Bowden (1973) have shown the attraction of tropical Pyrrhocoridae, especially Dysdercas, to light-traps. Four Jermy light-traps (see Duviard, 1972) with Philips HPL 125 bulbs were used. 190 DOMINIQUE DUVIARD

Three traps (Lamto, Foro-foro, Ferkessedougou) were operated twice a week (from 18.30 to 06.30 h G.M.T.=local time), while the trap in Adiopodoumé was operated every nigh,t from 18.30 to 06.30 h. The traps were worked for 41 months (September 1970-January 1974) in Foro-foro; 22 months (April 1972-January 1974) in Adio- podoume 19 months (July 1972-January 1974) in Lamto; and 13 months (January 1973-January 197 in Ferkessedougou. During the last 13 months, all of the traps were working.

U r 6. 2 4. 3' I I I 1 I

i

Fig. 2.-The four light-trapping sites in Ivory Coast plotted on the vegetational map. Grid area=ever- green rain forest; striped area=semi-deciduous rain forest; lightly-dotted area=forest-savanna mosaic; medium-dotted area=sub-Sudanian savanna; heavily- dotted area=Sudanian savanna. Simplified from Guillaumet (1971).

Trap Zocafions (Fig. 1 0nd 2) Four trap locations were selected, ranging from the south to the north of the country. These were : Centre ORSTOM d'ddiopodounzé (5"19"; 4"08W, 25 m elevation).--Sited on the balcony of the Entomology Department, the trap covered a forest clearing where the laboratories are established. Natural environment is secondary evergreen rain forest. The climate is of the D 1 type, described by Eldin (1971), with four seasons; long dry season lasting 3-4 months, with 0-1 month of harmattan; 1600 to 2500 mm of rain; 1800 to 2200 h of insolation; mean monthly temperature from 21 to 33°C (annual mean 2627°C). Climatic datal from July 1972 to January 1974 are presented in Fig. 3. Station d'Ecologie Tropicale de Lamto (6"13'N; 5"02'W; 104 m elevation).-The trap I was sited in savanna 150 m from the boundary of the gallery forest of the Bandama I

1 Data provided by Laboratoire' de Bioclimatologie, ORSTOM, Adiopodoumé.

c

..-. MIGRATIONS OF DYSDERCUS IN WEST AFRICA 191

J IIIAUI JASONOI 197J m FERKESSEDOUMU FORO FORO

LAMTO ADIOPODOUME.

Fig. 3.-Climatic data for the four light-trapping sites during the trappiag period. Each graph shows maximum, mean and minimum temperatures (as monthly means, C); monthly sunshine (hoursjmonth); monthly rainfall (mm) (in histogram), and calculated monthly potential evapo-transpiration (mm), (broken line). river. The locality is typical of the forest-savanna mosiac. The climate belongs to rhe C 2 type, described by Eldin (1971), with four seasons; long dry season lasting 4-5 months, with 15 days to two months of harmattan; 1200 to 1800 mm of rain; 1800 to 2000 h of annual insolation; mean monthly #temperature from 19 to 33°C (annual mean 25-28°C). Climatic data2 from July 1972 to January 1974 are presented in Fig. 3. Ferme I.R.C.T. de Foro-foro (7"jS'N; 5"OlW; 290 m eIevation).--The trap was sited in tree savanna, on the upper pant of a slope, 300 m from a gallery forestJ close to an experimental field (0.5 ha) cultivated with alternate rows of cotton and maize. Except for this small plot, the natural environment is nat disturbed and consists of tree savanna crossed by semi-deciduous gallery-forest. The area has a mixture of Guinean and Sudanian species. The climate belongs to the C 1 type, described by Eldin (1971), with two or four seasons; long dry season lasting 5-6 months with 1-3 months of harmattan; 1100 to 1600 mm of rain; 1800 to 2300 h of annual insolation;

Data provided by M. J. L. Tournier, Station de Géophysique, Lamto. .' 192 DOMINIQUE DUVIARD

mean monthly temperature from 19 to 34°C ('annual mean 25-28°C). Climatic data 3 from JuIy 1972 to January 1974 are presented in Fig. 3. Station I.R.A.T. de Ferkessedolrgou (9"35"; 5" 14'W; 325 m elevation).-The trap was sited in the courtyard of *theStation. The region is crossed by the boundary between sub-Sudanian and Sudanian savannas, $but is an area of intensive farming, and the natural environment is mostly eliminated. The climate $belongsto the B type, described by Eldin (1971), with two seasons; long dry season lasting 7-8 months, with 3-5 months of harmattan; 1100 to 1700 mm of rain; 2200 to 2700 h of annual insolation; mean < monthly temperature from 16 to 36°C (annual mean 2647°C). Climatic data4 from 4 January 1973 to January 1974 are presented in Fig. 3. f

Species of,Dysdercus trapped t Five species were caught: Dysdercus fasciatus Signoret, 1861; D. haemorrhoidaalis Signoret, 1858; D. I melaitoderes Karsch, 1892; D. sirperstitiosus (Fabricius, 1775); and D. voelkeri 11 Schmidt, 1932. A long-standing confusion of D. superstitiosus and D. voelkeri, under i the name superstitiosus, was clarified by Pierrad, 1967.) In 95 trap months 18 693 bugs were caught. Numbers, by species and locality, are given in Table II. The species fasciatus and superstitiosus are represented by single specimens; haeiiiorrhoidalis by 212 specimens, iiielaizoderes by 88S2 specimens and voelkeri 'by 9597 specimens. As shown by Table II and Fig. 4, the specific composition of the Dysdercus population in a given area, as observed by light-trapping, varies with latitude. D. nielanoderes is a forest species, seldom caught in the savannas of the

TABLEII. Numbers of Dysdercus caught in light-traps at four locations in the Ivory Coast 2 $ e 2 5 C 4 d 4 Ferkessedougou (Jan. 73-Jan. 74; d: 1298 76 1 13 months, 2 nightslweek) O: 1504 84 - Total: 2802 160 1 &IO: 0.86 0.90 - Foro-foro (Sept. 70-Jan. 74; 8: 2508 381 98 41 months, 2 nightslweek) 9: 3303 796 90 Total: 581 1 I177 188 ,e?/?: 0.75 0.47 I .O8 Lamto (July 72-Jan. 74; 8: 199 2840 IO 19 months, 2 nights/week) o: 175 2767 8 Total: 374 5706 18 81% 1-13 1-02 1.25 Adiopodoumé (May 72-Jan. 74; 8: 327 719 22 months, 7 nightslweek) 0: 283 1120 ; Total: 610 1839 5 81% 1.15 0.64 1-50 Overall totals: 9597 8882 212 i

3 Data provided by Laboratoire d'Entomologie agricole, ORSTOM, Adiopodoumé. 4 Data provided by Station Centrale de I'I.R.A.T., Bouaké. MIGRATIONS OF DYSDERCUS IN WEST AFRICA 193

LN LN t t Dv

613

5’19

100 200 300 CM 5 10 5 20 CM LN f

too 200 300 CM

Fig. 4.-Variation with Latitude North (LN) of mean monthly catches (CM) at the four light-trapping sites. Catches for Adiopodoumé have been calculated on the two nights per week basis used at the other three stations. DV=Dysdercus voelkeri, (R+ O, red-torange; J, yellow specimens); Dh=D. haemorrhoidalis; Dm=D. melanoderes.

northern Ivory Coast; D. haemorrhoidalis is mostly present in the savannas of the central part of the country and D. voetkeri is more numerous in savanna than in forest. There is a similar difference in Ghana between the distributions of melanoderes and . voefkeri (Leston, D., personal communication).

Seasonal variation in migratory activity Fortnightly totals of light-trap catches for the last 18 months (13 in ithe aase of Ferkessedougou) of the experiment are presented in Fig. 5. Cutches of D. voelken.-The phenology of this species differs according to locality, as previously observed (Duviard, 1973). In Ferkessedougou, following the decreasing flight aotivity observed during the short dull period (Gibbs & Leston, 1970) of heavy rains (July-August), catches increase during the ,inter-season (for climatic division of the year, see Fig. 6). A fmt peak is observed at that time (September). Catches decrease with decreasing rainfall until mid-November. Migratory activity is stopped during the continenkt1 dry season (December-January) and builds up again in February. A new peak is reached by the end of March, coinciding with athe first rains, and flight activity continues, although decreasing slowly, until the return of heavy rains. In Foro-foro, migrating D. voelkeri appear abruptly in October, just after the end of heavy rains, but are present for only one month. After an interruption of variable

... 194 DOMINIQUE DUVIARD

2. D. MELANODERES 1' 2, IItK1SStDOUOOU w

J?MAMJJAIONDJ

2

1

31 ADIOrODOUME

D. HCcMORRHOlDALIS

' ADIO~ODOUME b I 3 A'I'O'N'D'J'F M A'Y J'J'A S'0'N'D.J lm ,SI3 1974

l Fig. S.-Fortnightly catches of the three species of Dysdercus at the four light-trapping sites. I Histograms of log (n+l) transformation (scale from O to 3). 1 length (see Duviard, 1973) during the dry season, flight activity reaches its highest peak in February to April, during a warm sunny and .wet period, decreasing quickly afterwards. By the end of May, few specimens are caught. In Lamto, bugs appear in small number early in the dry season, between October and January, but important flight activity is not observed until March-April, during the warm, sunny and wet inter-season. During the rest of the year, catches are occasional or nonexistent. In Adiopodoumé, this species is caught by light-trap only during the dry season and the following inter-season (December to April). In the lower Ivory Coast, climate MIGRATIONS OF DYSDERCUS IN WEST AFRICA 195

is, by then, warm, sunny and wet. The highest peak of flight activity is observed in March. There are no catches at all during the other, dull and rainy, pant of the year. Catches of D. melanoderes. The phenology of this species also differs according to locality. In Adiopodoumé and Lamto, in the rain forest zone, flight occurs through- out the year, ,but with fluctuations. Less activity is observed during the period of very heavy rains, and in the rainless, dull, ' short dry season ' from June to September. In Foro-foro, catches are episodic. The largest ones occur during the inter-season (February to May) but there may #be some during the shont dry season (July to Sep- tember, according to years), or during the inter-season (October-November) preceding the dry season. There are no catches during the dry season or during the heavy rains. In Ferkessedougou, athe species is only caught, though in rather large numbers, during the initer-season (March to May) following the dry season. Catches of D. haemorrh0idalis.-The catches of this species were not numerous and numbers are insufficient to allow an analysis of its phenology. All that oan be said is that the bugs fly from November to May. In the four localities this covers the inter-seasons preceding or following the dry season, and, in the rain forest localities only, it covers the entire dry season.

Discussion Cornments on the observed meteorological features In this paper, a distinction has been made between the Inter-Tropical Convergence Zone (I.T.C.Z.) and the Inter-Tropical Front (I.T.F.). British authors always refer to the I.T.C.Z. and do not distinguish the I.T.F. separately. In the present case, such a distinction has been felt necessary. The zone of inter-tropical low pressures is swept by converging air flows from the anticyclonic areas situated at higher latitudes Those flows, comprising the I.T.C.Z., meet along a low-pressure axis, or convergence axis, and the surface of separation between the northeasterlies and the southeasterlies is the I.T.F. Thus the I.T.C.Z. covers a much larger area than the 1.T.F The confusion of I.T.C.Z. and I.T.F. leads to the idea, which is not always true, that convergence exists only near the I.T.F. Now, the I.T.F. itself is not always dis- playing strong convergence. Thus, in this paper, 1.T.F refers to the frontal surface alone, while I.T.C.Z. refers to the whole meteorological system. Such definitions of these terms, extracted from the work of Leroux (1972), are more precise and will prove quite useful in the following pages.

Flight activity, weather and survival The links existing between fight activity and types of seasonal weather at Foro-foro have been discussed in a previous paper (Duviard, 1973). The results of the present observations completely confirm this first interpretation of flight ecology in D. voelkeri. Whenever it is observed, the migratory activity is related to warm, wet and sunny conditions occumng in the climatic belts B and C (Fig. 6 and 7), as determined by the constitution of the I.T.C.Z. which, according to the evaluations of meteorologists (Eldin, 1971; Leroux, 1972), represents a !belt of 60-900 km width. In climatic zone A, north of the I.T.F., where the harmattan wind blows, the survival of D. voelkeri 4s only possible for a short he. If the relevant area has this type of weather for long periods (as in Bamako, Ferkessedougou and Kmno), D. voelkeri disappears, since temperature and humidity are not compatible with the survival of eggs, nymphs and adults, nor with locomotory activity. In climatic zone D, where heavy rainfall occurs, and in zone E, where humidity is stiI1 very high, despite lesser rains, the drowning of colonies of D. voelkeri and the development of a lethal fungal disease curtail populations of the species, which disappears (as at Adiopodoumé, Lamto, Foro-foro, Tafo, Kwadaso and Ibadan). A 196 DOMIXIQUE DUVIARD

I 10 hm -

5b. ------

Fig. 6.-Schematic structure of the Inter-Tropical Convergence Zone, showing the different climatic zones. A, continental dry season, with harmattan; B, oceanic dry season, with . predominant southwesterlies; C, interseason; D, rainy season; E, short dry season. The presence of bugs indicates the types of weather associated with migratory activity and potential breeding.

Fchruari I

August I .. Fig. 7.-LOCatiOn of the climatic zones A-E (see Fig. 6) at their extreme positions. The Ivory Coast is outlined. MIGRATIONS OF DYSDERCUS IN WEST AFRICA 197 . drier local climate allows further developmenit of populations, as observed rather late in the year (July) in the Accra plains by Fuseini (1972), but no food is available in this area when climatic belts D and E return in August-November. Thus, it is in the climatic zones B and C that migratory activity is concentrated, for only these climates favour the development of stainer populations as well as their flight adivity. It is there that the highest convective activity takes place in the atmo- sphere, 300 to 600 km south of Ithe I.T.F., according to Leroux (1972), who writes: ' calculations demonstrate the existence in Nigeria of hourly mean ascending movements of 5 km, and instant speeds of 25-30 kmh-l'. Such ascending movements may easily account for the upward transport of flying insects, thus allowing the bugs to escape their ' boundary layer ', recently defined by Taylor (1974).

Movements of the I.T.F. The existence of a close relationship between migratory activity and climatic belts associated with the structure of Ithe I.T.C.Z. may account for those sudden outbreaks which are one of the most conspicuous features in the ecology of cotton stainers. The I.T.F. is not a static structure and *this'climatic equator of Earbh' makes a double annual journey. The whole system moves between a southernmost limit, which at ground level may be defined as 7"30'N in February, and a northernmost limit at 22"N in August, at l@he longitude of the Ivory Coast (Fig. 7). This is a simplification of the reality, for the I.T.F. performs important diurnal movements as well as dispkcements of mean amplitude. According to Leroux (1972), the I.T.F. oscillates around its northernmost position between July and September, around its southernmost position between December and February. Except for these extreme situations, the I.T.F. moves over West Africa from south to north during six months, from north (to south during four months. The whole climatic system moves with the I.T.F.; but there is some distortion, as the declivity of the Front is steeper during its southernmost position than during its northernmost position. As a consequence, the climatic beIts are wider during the northern summer than during ,the northern winter. Besides these annual displacements, mean variations exist (Leroux, 1972) covering 5 to 8" of latitude (900 km) in periods of 3 to 15 days. Finally, the I.T.F. shows important diurnal oscillations, covering in 24 h distances ranging from 50 to 250 km and more, with inshnt speed above 50 kmh-l, correlated with diurnal modifications of atmospheric temperatures. According to Leroux (1972), the instant movements of the I.T.F. between 15.00 and 00.00 h are of the same general direction as the semi-annual migration. The existence of two main wind systems produces the I.T.F., since the movements of the I.T.C.Z. occur in the area of West Africa covered by either the northeasterlies or the southwesterly monsoon. When the I.T.F. oscillates around its northern limit, the whole of West Africa is swept by the wet southwesterlies of the monsoon. When the I.T.F. oscillates around its southern limist, West Africa presents two opposite areas: south of the I.T.F. and rcnder it, southwesterlies still prevail, while north of the I.T.F. and above it (and sometimes several hundreds of km south of the I.T.F. at ground level), the country is swept by the harmattan and northeasterly winds. This advance of the northeasterlies, dominating the monsoon, is more conspicuous during the southward movement of the I.T.F., ¡when its slope is a regular inclined plane, than during its northward movement, when the Front bulges under the pressure of the monsoon (Fig. 8). Besides these seasonal winds, correlated to the I.T.C.Z. itself, the oscillations of the I.T.F., by their amplitude and speed, draw along important atmospheric displacements. In particular, the air masses below the southern face of the Front are blocked there by this strong physical obstacle and do not mingle with the over-riding dry air; during the sudden southward movements of the T.T.F., these air masses are pushed back in the same direction (Roche, personal communication). 198 DOMblQUE DUWARD

A

UPPER WOBlHERLlES

C

B

W D 5 Fig. 8.-Migrations of Dysdercus spp. associated with atmospheric changes during the southward (A) and northward (l3) movements of the I.T.F. C, convective movements in the atmosphere associated with towering alto-cumulus clouds; LUS, low-level jet stream; broken-line arrow indicates reversal of air masses trapped under front; stippled arrows indicateaprobablemovements of migrating Dysdercus.

Annual displacements of D. melanoderes The case of D. melanoderes is relatively simple and aids interpretation of observa- tions on D. voelkeri. D. ïnelanoderes lives mainly in the rain forest zone, where dispersal flights occur all year round. The existence of two phases in the life of ;the insect, a migratory phase in young adults followed by a reproductive phase in older bugs, will be described elsewhere (Duviard, in preparation). The young adults leave their breeding habitats during their migratay phase. It seems that they escape their ' boundary layer' waibhout much difficulty. On several occasions we have observed catches of D. melanoderes in yellow water traps (which are i not a good sampling technique for this insect) hung at 12 m above ground level in an upper slope savanna in Lamlto (see Duviard ¿k Pollet, 1973, for a detailed account), more than 250 m higher than the nearest breeding zone, and thermal ascent or convection may have helped considerably the vertical ascent of these bugs. During the northward movement of the I.T.F., the winds are rather strong at low level in the first half of the night, once past the two hou" around sunset. At certain heights, the existence of a ' low level jet stream ' has been demonstrated by the experi- ments of G. W. Schaeffer (unpublished information) and is visible in the movements of clouds during clear nights, when much feebler winds may be observed at ground level. An insect that flies high enough to escape the boundary layer, either by active flight or with the help of convection, may thence be carried away by the southwesterlies (Fig. 8 B), even beyond its usual geographical area, as is shown by the peaks of catches in Foro-foro and in Ferkessedougou (Fig. 5). This northward movement occurs in February-May, during the passage of climatic belt C, when winds are of the same general direction as the movement of the I.T.F. In Foro-foro, where patches of semi-deciduous forest still exist, the species may live during the wet season, but disappas during the driest pant of the year and the habitats are newly colonised each year by young migrants from the south. MIGRATIONS OF DYSDERCUS IN WEST AFRICA 199

In Ferkessedougou, environmental conditions do not allow the survival of the species. The individuals that have been camed there by the monsoon winds cannot settle in an area where there is no semi-deciduous forest, and they therefore die. The migrants do not colonise, even temporarily, malvaceous hosts available in the savanna and no other flight period #has been observed, which suggests that there has not been any subsequent breeding.

Annual displacements of D. voelkeri The geographical area occupied by this species is restricted to its minimum dmen- sions at the two opposite periods of the year (Fig. 9). When the I.T.F. is at its southern

I 1 Fig. 9.-Extreme ranges of D. voelken' in August, during the northern summer (horizontal striped area between 500 mm isohyet and northern limit of climatic zone D) and in January, during the northern winter (vertical striped area between the ground limit of the I.T.F. and the coast). limit, the insect may live only in a narrow belt between the I.T.F. at ground level and the coast of the Atlantic (the width of the area is about 200 to 500 km,approximately covering the forest zone); and when the I.T.F. is at its northern limit, the insect may live only in a belt between the northernmost limit of climatic belt D (heavy rains}, as defined by Leroux (1972) and the 500 mm isohyet described as the northern limit of the species by Whitfield (1933), which is aIso the northern limit of malvaceous host plants (Fig. 1). The width of &is area is about 100 to 400 km. When the I.T.F. moves southwards, ithe geographkal area gets wider (Fig. 9) in its southern part, until the I.T.F. at ground level coincides with the 500 mm isohyet. Then the whole area, 600 to 900 km wide, moves southward until it is diminished by the advance of the climatic belts above the sea, The reverse phenomenon takes place during the northward migration of the I.T.F. This supposes that the insects use some sort of mechanism (for such displacement, whose magnitude between the coast and 500 mm isohyet vanes from 900 km (longitude 4"E) to 1250 km (longitude 8"W). In fact, the bugs, whatever the part of the geo- graphical area occupied at a given period, remain permanently in the climatic conditions of belts B and C (mostly C on the northern part of the area, mostly B on its southern part). They find there, as previously discussed, the best available conditions to breed and to migrate. Conditions in these belts allow the species to escape its boundary layer 200 DOMlNlQUE DUVIARD

with the help of convective movements, and flights within these belts do not show any particular orientation (see Schaeffer, 1972; Taylor, 1974). During the northward migration of the species, the mechanism is the same as that used by D. inelaiioderes, namely transport by predominant southwesterlies blowing regularly during the hours of migratory flights (from dawn to midnight, mostly-see Duviard, 1973), reinforced at low elevation by the ' low level jet stream ' (Fig. 8 B). Such a displacement explains sudden immigrations occurring at the end of the dry season, as in Foro-foro, Frekessedougou, Bamako and Kano. There is an apparent contradiction between the southward movement of the species during the same movement of the I.T.C.Z., and the existence, south of the I.T.F., of predominant southwesterlies. But a careful examination of the meteorological environ- ment helps to resolve the problem.

AUGUST /æH HARMATTAN E 9.w IE c--

*- *- OCTO BER //e- HARMATTAN 4 s.w ' /- -

JANUARY

-/- DwDm //

MARCH / /- c c

Fig. 10.-Modifications of the latitudinal ranges of D. voelkeri (Dv, dotted zone) and D. melanoderes @m, striped zone) associated with seasonal movements of the I.T.C.Z. The I.T.C.Z. is schematically represented and arrows indicate prevailing winds. The 500 mm isohyet i shows the approximate northern limit of D. voelkeri. i t j MIGRATIONS OF DYSDERCCJS IN WEST AFRICA 201

In the southbound climatic zone B, the first half of the night is windless (Duviard, 1977). During the daily oscillation of the I.T.F., (the atmospheric masses blocked under the Front may be swept 50 to 250 km southward, with an instant speed of 50 kmh-' (see above) and bugs flying in these air masses may be easily swept southward (Fig. 8 A). Convection may also carry the insects higher (as observed by Schaeffer, unpublished). In belts B and C the monsoon is not very deep during part of the year (maximum 1000 to 2000 m) and the insects may be taken into the upper northerlies and separated from the monsoon by the calm shift of the I.T.F. We know little about the possible height of flight of Dysdercus. During the early dry season of 1968, many D. voelkeri were seen in the upper grassland of Mount Nimba (1300 m and more) land they had probably been carried there by strong ascending winds, frequent at that period of the ya(Leclerc et al., 1955) on the slopes of this lonely ridge. Rainey (1963) made similar observations on the desert locust in the East African Highlands. In Uganda, espeoially the Westem Region (Ankole, Kigezi), dainers wre very common and cause much damage at elevations over 1200 m (Bowden, 1970). Also. stainers are common in West Nile Distriut, in Noh-West Uganda, at an altitude of 1200-1300 m (Bowden, personal communica- tion). Altitude per se does not seem (to affect the Occurrence of Dysderczis, but will be important only as it affects temperature, the limit of 18°C remaining the most important factor (Duviard, 1973). In Kenya, few stainers are caught in light-traps above 1200 m, probably due to the low temperature occurring there (Robertson, personal communication). If the transport of stainers by swept-back ,atmospheric masses seems to us, as well as to Roche (personal communication), 'the most probable explanation near the I.T.F. (climatic zone B), Dysdercris may, farther south (climatic zone C)be carJied high enough to ' catch ' the upper northerlies. Temperature there is perfectly consistent with possible flight for (thespecies (Leroux, 1972). Although one must keep in mind that the I.T.F. is extremely mobile and that conditions vary during the same night, as shown #by Haggis (1971}, Bowden & Gibbs (1973) and Duviard (1973), this southward movement accounts for the sudden immigra- tions occurring at the end of the rains, as observed in Foro-foro, Lamto, Adiopodoumé, Tafo, Kwadaso and Ibadan, and for their different timings. The Atlantic Ocean may not stop southward migrations of inseots. During December 1973, a strong surge of the harmattan carried the annually dispersing grasshopper Zonocerus variegoiiis (L.) far out to sea and masses of these insects were swept back by waves to the shore of the Ivory Coast, between Abidjan and Grand Bassam. This annual passage from one ecological area to the other may be compared to a genuine 'transhumance' (Fig. 10) and &is notion is acceptable in a continent where many species, as well as several human societies, ,have developed comparable ways of life, adapting themselves to severe alternation of ecologiml conditions. However, the return migration is not performed by the same individuals, as it is in the cases of the cattle egret or the Peuhl tribe. Each genevation of catton stainers accomplishes a short stage of the whole migration, invading progressively the newly available areas which remain suitable at least long enough for one generation to bred, as at Foro-foro (Duviard, 1972, 1973). That the newly formed adults go on with this obligatory transhumance is shown by the presence of distinct colour forms (p. 188). It is possible to estimate the average distance travelled in each generation. If we assume 8-9 generations a year, that is 4-5 for each single journey, each generation bas to travel 250 to 300 km at most, if migration is to be accomplished without difficulty. But populations must disphce as a whole, as shown by sudden outbreaks, their displace- ment being triggered by the establishment of adequate climatic conditions. For the insects leaving, for exampl, the northern part of an occupied area, to invade and OCCUPY the adjacent area further north, the whole journey is rio more than f to + of the whole range covered by the population of that generation. Each insect will then travel 100 to 150 km at most. (L 2283) B c .-.

302 DOìvl1?4lQUt DUVIARD

Taking account of the flight speed of the insect (10-12 kmh-l, Duviard, in prepara- tion), the ground speed of the migrating bugs will be the speed of the wind at the height of flight plus or minus khe flight speed, according to the direction of flight (downwind or not). The radar observations of Schaeffer (unpublished) give a range of 10-80 knh-', with an average of 30 kmh-l, for African night-flying insects. If insects fly for three consecutive hours, which seems to be the general case (Schaeffer, unpublished), and not a maximum for D. voelkeri (Duviard, 1973), a 90-km run is a possible nightly range. It would therefore take two to three consecutive nights, and six to nine hours of flisht to complete the necessary journey. Johnson (personal communication) has observed that D. intermedius Distant can fly for more than 25 consecutive hours in a tethered position, and D. peruvianus Guérin-Méneville is supposed to rtravel across a 50 km desert in Peru (Alza, 1959; Saldarriaga, 1959). Similar journeys then seem quite within the capability of D. voelkeri.

Conclusions Bowden (1973) and Rainey (1973) have recently reviewed the cases of association between regular, long-distance migrations of several insect species land the I.T.C.Z. in Africa. Best-known examples itre those of the desert locust, Schistocerca gregaria (Forskål) (Rainey, 1954, 1973) and of the African armyworm, Spodoptera en-einpta (Walker) (Brown et al., 1969). In those cases, which more particularly refer to East Africa, as in the examples given by Bowden & Gibbs (1973) in the Sudan, two facts are obvious: (i), the association of the flight aativity of insects with a definite type of weather, dependent on the general climatic system (I.T.C.Z.), and (iì), long-range dispbcements of insects favoured by winds associated with the climatic system itself. In West Africa, associations between migratory flights of insects and types of weather are known (Bowden, 1964, 1973; Gibbs & Leston, 1970) and long-range displacements have been studied in locusts (Bauen, 1967; Geubier, 1966). Wthile there is no direct evidence of flight by Dysdercus coincident with the passages of the I.T.F. (the whole flight activity of the bug always taking place on the southern, wet, side of the I.T.C.Z.), the movements of the latter are the only logical explanation of what is taking place. The case of Dysdercus spp. may now be added to the list of long-range migrapt insects in that part of the world.

Acknowledgements The author is indebted to sthe Institut de Recherches du Coton et des Textiles exotiques (I.R.C.T.) for providing field work kcilities, and more especially to M. Gramain, who took charge lof the light-trap in Ferkessedougou and to Dr R. VuaWoux, Institut d'Ecologie Tropicale de l'université d'Abidjan, who took charge of the light-trap in Lamto, Professeur Bergerard and Professeur Carayon kindly gave valuable criticism and help, but the author is especially indebted to Dr J. Bowden whose help was essential for the elaboration of this paper. '

References ALZA,D. (1959). Normas y conocimientos preliminares en el estudio de la migración del arrebiatado (Dysdercus peruvianus Guérin) en le Costa Central.-Revia peru. Ent. agric. 2, 86-91. BALLARD,E. & EVANS,M. G. (1928). Dysdercus sidae Montr. in Queensland.-Bull. ent. Res. 18, 405-432. BAITEN,A. (1967). Seasonal movements of swarms of Locusta migratoria niigratorioides (R. & E) ín Westem Africa in 1928 to 1931.-Bull. ent. Res. 57, 357-380. BEBBINGTON,A. G. & ALLAN,W. (1936). The food cycle of Dysdercus fasciatus in Acacia savannah in Northem Rhodesia.-Bull. ent. Res. 27,237-249. BERRIDGE,M. J. (1965). The physiology of excretion in the cotton stainer, Dysdercus fasciatus Signoret. I. Anatomy, water excretion and osmoregulation.-J. exp. Biol. 43, 511-521.

-, r MlGRATIONS OF DYSDERCUS IN WEST AFRICA 203

BOWDEN,J. (1964). The Sphingidae (Lepidoptera) of Kwadaso, Ghana, with special reference to their phenology, the influence of moonlight on activity and the effect of weather conditions on abundance and activity.-Ph.D. thesis, Univ. Bristol. BOWDEN,J. (1970). Cotton pests. pp. 179-180 in Jameson, J. D. (Ed.). Agriculture in Uganda- 2nd edn, 395 pp. London, Oxford University Press. BOWDEV, J. (1973). Migration OP pests in thc tropics.--Medede[ingen Fakulteif Landborcw- wetenschappen, Gent. 38,785-796. BOWDEN, J. & GIBBS,D. G. (1973). Light-trap and suction-trap catches of insects in the northern Gezira, Sudan, in the season of southward movement of the Inter Tropical Front.-Bull. ent. Res. 62, 571-596. BIIOWN,E. S., BETC~, E. & RAINEY,R. C. (1969). Seasonal changes in the distribution of the African armyworm, Spodoptera exempta (Wlk.) (Lep., Noctuidae), with special reference to East Africa.-Bull. ent. Res. 58, 661-728. DELATTRE,E. (1950). A propos des Dysdercus du cotonnier.-Coton Fibr. trop. 5, 93-94. DINGLE,H. (1972). Migration strategies of insects.-Science, N.Y. 175, 1327-1335. DINGLE,H. (1974). The experimental analysis of migration and life-history strategies in insects. pp. 329-342 in Barton Browne, L. (Ed.) Experimental analysis of insect behaviour.- 366 pp. Berlin, Springer-Verlag. DINGLE,H. & ARORA,G. (1973). Experimental studies of migration in bugs of the genus Dysdercus.-Oekologia 12, 119-140. DUVIARD,D. (1972). Les vols migratoires de Dysdercrrs voelkeri Schmidt (Hemiptera : Pyrrhocoridae) en Côte d’Ivoire. I. Le rythme endogène fondamental.-Coton Fibr. trop. 27, 379-388. DUVIARD,D. (1973). Les vols migratoires de Dysdercus voelkeri Schmidt (Hemiptera : Pyrrhocoridae) en Côte d’Ivoire. II. Les rythmes exogènes.-Coton Fibr. trop. 28, 239-252. DUVIARD,D. (1977). Ecologie de Dysdercus voelkeri en Afrique Occidentale : Migration et colonisation des nouveaux habitats.-Doctorate thesis, Université de Paris Sud, France, multigr., 165 pp- DUVIARD,D. & POLLET,A. (1973). Spatial and seasonal distribution of Diptera, Homoptera and Hymenoptera in a moist shrub savanna. Ecological behaviour of winged insect populations in the savannas of Ivory Coast. I.-Oikos 24,42-57. EDWARDS,F. J. (1969~). Development and histology of the indirect flight muscles in Dysdercus intermedius.-J. Insect Physiol. 15, 1591-1599. EDWARDS,F. J. (19696). Environmental control of flight muscle histolysis in the bug Dysdercus intermedius.-J. Insect Physiol. 16,2013-2020. EDWARDS,F. J. (1970). Endocrine control of flight muscles histolysis in Dysdercus intermedius. -J. Insect Physiol. 16,2027-203 1. ELDIN,M. (1971). Le climat. pp. 73-108 in Avenard, J. M. et al. Le milieu naturel de Côte d’Ivoire.-391 pp. Paris, Mémoires O.R.S.T.O.M. no. 50. FUSEINI,B. A. (1972). The biology of cotton stainers, Dysdercrrs spp. (Hemiptera : Pyrrho- coridae) in Southern Ghana.-MSc. thesis, Univ. Ghana, Legon. GSICIJET, P. F. (1956). Quelques facteurs de réduction naturelle dans une population de Dysdercus superstitiosus Fab. (Hemiptera, Pyrrhocoridae).-Revue Path. vég. Ent. agric. Fr. 35,2749. GERBIER,M. E. (1966). Analyse des relations entre la météorologie et le déplacement des essaims de criquets pélerins grégaires dans l’aire occidentale d’invasion.-F.A.O. publication no. UNSF/DL/RFS/4. G~Bs,D. G. & LESTON,D. (1970). Insect phenology in a forest cocoa-farm locality in West Africa.-J. appl. Ecol. 7,519-548. GOLDING,F. D. (1928). The control of cotton stainers in Southern Nigeïia.-Emp. Cott. Grow. Rev. 5,128-133. GUILLAUMET,J. L. (1971). Carte de la Végétation, in Atlas de la Côte d’Ivoire.-O.R.S.T.O.M., Abidjan, Map no. A6q. HAGGIS,M. J. (1971). Light-trap catches of Spodoptera exeinpta (Walk.) in relation to wind direction.-E. Afr. agric. For. J. 37, 100-108. HALFENBERG,R. & STEIN,G. (1971). Ober das rote Pigment des Baumwollwanze Dysdercus intermedius Distant.-2. Naturf. Ser. B. 26, 71. JOHNSON,C. G. (1969). Migration and dispersal of insects by flight.-763 pp. London, Methuen. V’C *

. 2001 DOMINIQUE DUVIARD ,Y KEAY, R. W. J. (1959). Vegetation map of Africa south of the Tropic of Cancer.-24 pp. 4- fldg. map. London, Oxford University Press. LECLERC,J. C., RIcHARD-h/loLLm, J., LAMOTTE,M. & TOURNIER,J. L. (1965). La Réserve naturelle du Mont Nimba. III. La chaine du Nimba, Essai géographique.-I.F.A.N., Dakar, 281 pp. LEROUX,M. (1972). La dynamique des precipitations en Afrique Occidentale.-Ph.D. thesis, Univ. Dakar. PEARSON,E. O. (1958). The insect pests of cotton in tropical Africa.-355 pp. London, Empire Cotton Growing Corporation & Commonwealth Institute of Entomology. PIERRARD,G. (1967). Sur la confusion de deux espèces du genre Dysdercus Guérin-Méneville (Pyrrhocoridae) : D. supersfitiosus Fabricius et D. Völkerì Schmidt sous le taxon superstitiosus.-Coton Fibr. trop. 22,421-424. PIERRAIZD,G. (1972). Le contrôle de Dysdercus voelkeri Schmidt défini par l’acquisition de connaissances de Ia biologie de l’insecte et de ses dégâts.-Doctorate thesis, Faculté des Sciences Agronomiques de l’Etat, Gembloux, Belgium, multigr., 297 pp. RAINEYR. C. (1954). Meteorology and the migration of desert locusts.-Tec. Notes WZd. met. Org. RAINEY,R. C. (1973). Airborne pests and the atmospheric environment.-Weather, Lond. 28, 224-239. SALDARRIAGA,Y. (1959). Contribución al estudio de las migraciones del arrebiatad0.-Revtu peru. Ent. agrié. 2, 105-108. SOUTHWOOD,T. R. E. (1960). The flight activity of .-Trms. R. ent. Soc. Lond. 112, 173-220. SOUTHWOOD,T. R. E. (1961). Notes on light trap catches of Heteroptera made in the tropics. -Entomologist’s mon. Mag. 96, 114-117. SOUTHWOOD,T. R. E. (1962). Migration of terrestrial in relation to habitat.- Biol. Rev. 37, 171-214. TAYLOR,L. R. (1974). Insect migration, flight periodicity and the boundary layer.-]. Anim. Ecol. 43, 225-238. WHITFIELD,F. G. S. (1933). The bionomics and control of Dysdercus (Hemiptera) in the Sudan.-Bull. enf. Res. 24,301-313. YOUDEOWEI,A. (1967). The reactions of Dysdercus intermedius (Heteroptera, Pyrrhocoridae) to moisture, with special reference to aggregation.-Ent. exp. uppl. 10, 194-210.

(Received 7 June 1976) .-

0 Commonwealth Agricultural Bureaux, 1977